Tamkang Journal of Science and Engineering, Vol. 13, No. 4, pp.

395-404 (2010) 395

Intelligent Safety Warning and Alert System for

Car Driving
Shih-Nan Lu1, Hsien-Wei Tseng2, Yang-Han Lee1*, Yih-Guang Jan1 and Wei-Chen Lee1
1
Department of Electrical Engineering, Tamkang University,
Tamsui, Taiwan 251, R.O.C.
2
Department of Computer and Communication Engineering, De Lin Institute of Technology,
Tucheng, Taiwan 380, R.O.C.

Abstract
The new vehicle performance has been continuously improved and the study results relating to
the safety of car driving have also been continuously reported and demonstrated, it is trying to find a
balance point between the development of vehicle speed limit and the protection of drivers safety. In
the current study and development of various products, no matter it is in the enforcement of vision
system, radar detection or the tracing and control it is always asking the driver to watch or handle the
possible issues after the occurrence of accidents. In this paper we try to develop a system to provide the
prior to accident information to the vehicle control unit so that it enables the vehicle to prevent the
happening of accident. During the vehicle movements the system will continuously record the
vehicles moving status and conditions so that the record will provide the decision basis in the accident
investigation if it unfortunately happens the fatal accident.

Key Words: Intelligent Safety, Vehicle Control Unit, Warning and Alert System, Automation,
Car Driving

1. Introduction
When it turned into twentieth century various kinds
of vehicles have been introduced to provide convenience
in human daily life and due to the developments of new
technologies it makes the vehicle running fast and ac-
celerating easily. It on the contrary brings in some pro-
blems such as the happenings of accidents due to drivers
fatigue after a long journey of travel, the shortage of
parking spaces, the bad visibility in night or the driving
in heavy rain days etc. The accidents sometimes bring in
the fatality of human lives and loss of properties, there-
fore many measures to enforce the safety driving and the
development of models to monitor drivers behaviors
have been proposed and brought in many promising re-
sults. From a study conducted by Mercedes Benz it re-
ports that if it can get an extra 0.5 seconds in the warning
period it will avoid 60% of the overrun accidents and it
*Corresponding author. E-mail: yhlee@ee.tku.edu.tw
will reach 90% if 1.5 seconds is added into the warning
period. We introduce in the following some safety warn-
ing systems [1-6] that are currently available in the mar-
ket and they are technically matured:
1.1 Sleepiness Warning System
It reports in statistics that the fatality rate increases
by four times when the driver is sleepiness in driving
therefore many researches are focused on finding the re-
lationship between the sleepiness with the drivers eye-
lids width, the visibility of the pupil, the motion of the
head etc. Industrial Technology Research Institute
(ITRE) has even used the ultra wide bandwidth (UWB)
technique to integrate low power pulsed electromagnetic
(EM) waves to precisely measure the drivers physiolog-
ical signals such as his heartbeat and respiration etc. It is
also through the development of various system algo-
rithms to detect drivers psychic status. In all these tech-
nology developments its main purpose is trying to emit
warning signal to awake the driver before his falling into
396 Shih-Nan Lu et al.
sleepiness and to avoid any possible accidents.
1.2 Low-Speed Following Mode (LSM) System
The LSM exploits mini-meter wave radar to detect
the acceleration, deceleration and stopping of the front
car to estimate the distance away from the front car,
meanwhile he also controls the brake and the fuel sys-
tems to maintain his car within the safety range. If the
front car experiences abnormal condition the system will
blow alarm sound to warn the driver.
1.3 Lane Keeping Assist System (LKA)
The LKA uses camera to monitor the passing or di-
viding lane in the front and the system will give warning
signal if the driver crosses the passing lane or enters into
the opposite direction passing lane without giving or gi-
ving improper direction signal.
1.4 Intelligent Night Vision System (INVS)
In driving at night or in heavy rain the drivers visi-
ble range will be restricted and also the light illuminating
range will be limited to prevent the driver from clearly
noticing any pedestrian along the roadside or the person
is in repairing his out of condition car the INVS will dis-
play the front road condition on the LCD screen of the
car audio system by using infrared camera to monitor the
road condition to present to the driver a complete know-
ledge of the front road condition to greatly reduce the
possible occurrences of accidents.
1.5 Adaptive Front-Lighting System (AFS)
The AFS catches the cars moving direction by using
the direction sensor implemented in the wheel and then
using the units control motor to rotate the headlight and
with the installation of High Intensity Discharge (HID)
Xenon light to extend the light illuminating range so as
to enable the pedestrians and the obstructions in the ve-
hicle direction to be visible to the driver.
1.6 Video Parking Assist System
This system uses camera system to extract the images
of the car in the back and side directions and in its coordi-
nating with the collision avoidance radar system and the
dynamic auxiliary linear system to inform the driver of the
wheel rotation direction and angle to assist the driver to
complete the reversing or parking action. Currently it has
a new system, in coordinating with the wheel rotating
mechanism, without requiring the direct involvement of
the driver to automatically assist the driver to reverse the
car into the garage or to park the car along the road side.
It has other systems, just mention here for reference,
such as Adaptive Cruise Control (ACC), Electronic Sta-
bility Control (ESC), Emergency Brake Assist (EBA),
Blind Spot Warning etc. each system has its own specific
functions and will not discuss them further [7,8].
In the design of safety vehicle driving system, it needs
to consider and include the following four processes:
(1) Normal condition - to assist the driver to lower his
driving burden and to provide him with visual and
driving assistances.
(2) Prior to accidence - to prevent the occurrence of acci-
dent and to possibly provide the actions of collision
avoidance and warning.
(3) During the accidence - to reduce the damage and the
collection of accident data and information.
(4) After accidence - to report the accidence and to pro-
vide emergency treatment.
In the above-mentioned research themes, they usu-
ally emphasize and confine in the normal and prior to ac-
cidence processes it seldom covers the information and
data collections and the emergency treatment during the
accidence and the accident report after the accidence.
The main theme we will consider in this paper is the
monitoring of the vehicle behavior model prior to the
happen of accidence, to warn the driver when abnormal
and dangerous condition might happen and meanwhile it
will record the vehicle acceleration, deceleration and its
moving direction and finally to provide sufficient infor-
mation or data for the investigator when he is in the in-
vestigation of the accident case.
The organization of this paper is in the following.
We will introduce the current development of vehicle
safety system in section 1. In the second section we will
briefly introduce the most commonly implemented in-
formation transmission system in the current vehicles.
We will then introduce the design concept and the test
platform of our proposed vehicle monitoring system in
section 3. In section 4 it will present, discuss the data we
collect and the results of analyzing these collected data.
2. Introduction of Vehicle Transmission System
It currently has more than few tens of electronic in-
struments installed in the vehicle, but at the initial design
stage almost every instrument is just designed to inde-
 Intelligent Safety Warning and Alert System for Car Driving
pendently perform its own function without coordinating
and accommodating with others so that the vehicle instru-
ments are just providing the base layer data of the vehi-
cle condition such as the vehicle speed, engines torque,
the remaining in the gas tank and the radiators tempera-
ture etc. The vehicles audio-video system just provides
the audio or video amusement and could not display or
provide other conditions of the vehicle. This independent
situation has quickly been drawn attention and every
world car manufacture has proposed its own ideas of how
to inter-connect instruments, basically they possess the
characteristics of separating the high speed instrument
from the slow speed instrument, serial bus structure and
the capability to resist the high random noise. We give a
brief description of each instrument in the following:
2.1 MOST bus (Media Oriented Systems Transport)
The MOST bus system was agreed upon and set up
by some giant car manufactures and cars equipments
providers in 1997 it uses the structure of Ring, Star or
Daisy-chain to build up a point-to-point optical network
system with data rate of 22.5 Mb/s that it is high enough
to convey the compressed videos. Plastic optical fiber is
the main transmission media for the optical network to
provide a wide bandwidth, low interference data trans-
mission network. The Germany car manufacture BMW
integrates this optical network with its own developed
ByteFlight system to form the Intelligent Safety Integra-
tion System (ISIS). This ISIS system has been imple-
mented in its 7 series sedans to provide a safety monitor-
ing function for the whole vehicle.
Due to the high transmission speed of the MOST bus,
the low radiation interference to its surroundings, the
lightweight of the transmission material (it is estimated
that its weight is around 1/20 of the current copper wire),
the price dropping of the optical material and the easier
installation and connection of the optical network etc., it is
anticipated that the MOST bus system will become the
dominant system in the vehicle bus system [9-11].
2.2 J1850 bus
J1850 bus was designed in 1994, its purpose was to
apply the functions of self-diagnosis and information
sharing etc. within the transportation vehicle (such as
car, airplane etc.) In J18050 bus it basically has two infor-
mation transmission modes, the first mode is a 41.6 Kbps
dual-line differential PWM transmission mode, the other
is a 10.4 Kbps single line VPM transmission mode. Cur-
397
rently the majority of cars uses the single line transmission
mode its transmission distance can reach 35 meters [11,12].
2.3 CAN bus (Controller Area Network)
Within the transportation vehicle it can install one
set or many sets of CAN buses. Low speed CAN bus can
be used in the regular information transmission and con-
trol while the high speed CAN bus can be used in the
handling and management of engine condition and in the
control of vehicle brake system. The CAN bus is the
most popular bus system used in all vehicles, almost all
information within each vehicle can be communicated
and conveyed through this bus system, but each car
manufacture has different definition for its information
content, it is not easy to understand the content of the in-
formation transmitted in the bus of other vehicles if the
information definitions in the bus system are not trans-
parent between car manufactures [11,13,14].
2.4 LIN bus (Local Interconnect Network)
This bus system is used within the transportation
vehicle itself, it is a series bus to communicate between
each detecting/monitoring instrument and electronic
switch, its highest transmission rate is 19200 bauds for a
maximum transmission distance of 40 meters [11].
In addition to above discussed bus structures that are
frequently implemented in the vehicles they are many
other designs and structures of bus systems such as MI,
DSI, OBDII, Flex Ray, D2B, SMART wire X, IDB-1394,
IE bus, Intelli Bus, BST bus, MM bus and J1708 bus etc.
3. Design Concept and the Composition of the
Design Platform
3.1 The Composition of the Design Platform
This design platform consists of a microprocessor
and an acceleration sensor as shown in Figure 1 is a test
board of an acceleration sensor. The analog signal output
from the acceleration sensor is amplified by a two-stage
amplifier and is sent to the design platform, the system is
completed by connecting this platform with a computer.
This test board is also integrating with an optical receiver
and many sets of analog/digital converters for use with
other applications.
3.2 Procedures in Data Extraction and Data
Analysis
It uses vibrator to test the acceleration of the design
398 Shih-Nan Lu et al.
platform, the design platform is vertically fixed at the vi-
brator platform due to the vibrator is repeatedly vibrating
in the up/down directions. We made four random tests
and in each test we applied 1G ~ 4G vibrating force test
in four directions of the design platform. As shown in
Figures 2 and 3 are a series of resulting vibrating wave-
forms in the tests. The vibrator delivers its sensed output
in each dimension of the 2D test, but we just select one
set of data from the x-axis and from the y-axis for analy-
sis. In the test driving condition when it happens to have
acceleration impulse effect it will happen only in the x-
direction (forward-backward) with an abrupt and high-
slope data changes while in actual driving situation it
will also happen with data variation in the y-axis (left-
right). The data obtained from these four random tests
are almost the same therefore we can conclude that the
Figure 1. Test board of acceleration sensor.
Figure 2. Test data in the x-axis.
data obtained from the design platform are reproducible
and we can use any set of data in the analysis.
After analyzing the test data, obtained from the de-
sign platform, we have the data distributions for the
1G~5G vibrating force tests as shown in Figure 4.
3.3 The Design of Acceleration Impulse
Pre-Warning System and Its System Flow Chart
When a vehicle is supposed to be in slow motion or in
stopping situation but due to the malfunction of the fuel
supply system or due to the human mistake the vehicle en-
gine is accelerating to a high speed in a very short duration
and resulting in the vehicle damage and the fatalities of
the driver and passengers. It intends to design a pre-warn-
ing system to prevent the happening of the acceleration
impulse in its early forming stage and to inform or to warn
on time to the fuel supply system, the electric system or
the brake system to perform appropriate action to prevent
the occurrence of this kind of accidence.
The acceleration impulse pre-warning system is ba-
sically an automatic data analysis system that is set up
from integrating the results by analyzing various test
data collected from the test. When the design platform
extracts a new data it will analyze the data and from its
analyzed result to determine how many Gs of the vehi-
cles acceleration has and to display the result on the
monitor. In the design of our pre-warning system we as-
sume that when the acceleration exceeds 3G it will have
acceleration impulse effect, therefore we will mark with
red symbol under the main picture when the data is
 Intelligent Safety Warning and Alert System for Car Driving 399

Figure 3. Test data in the y-axis.

Figure 4. Data distribution resulting from 1G~5G vibrating force tests.

higher than 3G to signify it is in the dangerous stage.
Furthermore in the actual test we found that when the ac-
celeration exceeds 5G this vibrating force is far more
than human being can be safely driving or seating in the
vehicle we therefore do not extract any data when the ac-
celeration force is more than 5G and the system design
limit is also set at 5G. The system flow chart as shown in
Figures 5 and 6.
3.4 System Design for Monitoring the Driving
Behavior and Its System Flow Chart
Except for few buses and large-size touring buses, al-
most all compact vehicles are not equipping with any
driving behavior monitoring system therefore it is impos-
sible to trace drivers driving behavior for those cars dri-
vers. Consequently they do not have any pre-warning
and verification data available during the time prior to-,
during and after the accident. The driving behavior moni-
toring system is not only to record the vehicle running con-
ditions for certain time interval but also gives warning sig-
nal to the driver before occurring any dangerous condition
and to advise the driver to correct his driving behavior.
When a car is running on the road the driver will al-
ways have certain preparatory action when he is going to
perform any change, which can be modeled by an as-
sociated identifiable model, for example, when he de-
cides to make right or left turn he will normally reduce
his speed and enters into the turning lane he will then ac-
celerate after he enters into 1/2 or 1/3 of the turning lane.
In another example when a driver is driving in high
speed and is going to make an emergent stop he will usu-
ally take a few stages to break or stopping the car instead
of braking the car all the way in only one step from the
considerations of the safeties of the passengers and pos-
400 Shih-Nan Lu et al.
Figure 5. Flow chart of acceleration impulse pre-warning
system (1).
sible material carried in the vehicle.
Based on the experience of driving a car and col-
lected statistical data in normal driving situation we can
analyze and develop some abnormal driving behaviors
such as when the driver is not fully awake or he is uncon-
sciously drunk that he is driving the car diverting toward
to one side or could not keep the car running in a straight
line due to he could not make correct decision or to deter-
mine the proper distance from the front car, or he could
not clearly identify in front road condition or he has
fallen asleep and even could not control the car. In this
situation he is driving the car toward to right or left but
he will not have any braking action and in the worst si-
tuation he is speeding while the car is constantly making
changes in right or left turns.
The intelligent driving monitoring system is an auto-
Figure 6. Flow chart of acceleration impulse pre-warning
system (2).
matic data analysis and simulation system. Since the be-
havior model used in the monitoring of driving behavior
is rather complicated and it also needs to analyze data
taken simultaneously in the x-axis and y-axis directions,
in this paper we design and implement a simplified driv-
ing behavior monitoring system to analyze only one x-
axis and one y-axis data to determine the drivers behav-
ior. We will need in our future task to enforce this intro-
ductory system to consider not only one data point but to
monitor the variation of data in an interval of time and
then to make the decision from these integrated data in-
formation. It may also need to include some fuzzy pro-
cess so that to accomplish a more delicate identification
results. The system flow chart as shown in Figures 7 and 8.
4. Test Results and Its Distinguish from
Other Systems
4.1 Test Results of Acceleration Impulse
Pre-Warning System
When a vehicle endures acceleration impulse force it
 Intelligent Safety Warning and Alert System for Car Driving
Figure 7. Flow chart of driving behavior monitoring system
(1).
is almost for sure it will be abruptly moving toward for-
ward or backward, the engine reinforces it to continu-
ously rotate and is moving faster. Consequently the de-
sign principle of our pre-warning system is to use the ac-
celeration sensor to measure and analyze the amount the
car has moved forward or backward; make its decision
based on the analyzed result. The test platform is in-
stalled on the car and it will measure the actual output of
the acceleration sensor when the car suddenly acceler-
ates. After analyzing the measured data it is found that
when the sudden acceleration force has reached around
2G the car is still controllable therefore we assume that it
will incur acceleration impulse effect when it reaches 3G
acceleration force and we will use 3G as the threshold to
issue a warning signal.
We input the measured data into the acceleration im-
pulse detection system; when the input data has force
higher than 3G it will mark on the figure a red warning la-
401
Figure 8. Flow chart of driving behavior monitoring system
(2).
bel. When the system is actually implemented in the vehi-
cle this warning label and control command will flow
through the data bus directly into its relevant systems to
provide a timely safety protection measure or signal.
4.2 Test Results of the Driving Behavior
Monitoring System
In the driving behavior monitoring test the data
taken from the x-axis of the design platform is relating to
the vehicle acceleration and braking decision while the
data taken from the y-axis axis is relating to the decision
to make right or left turn. In Figures 9~11 they analyze
and display the results of many models of normal or ab-
normal driving behaviors. In the figure the label AD0
denotes the forward or backward movement, while AD1
represents the left or right direction movement. For an
illustrative example consider the third section of Figure
9, it initially runs the test with high speed but with
non-braking action it happens that the speed is too high
402 Shih-Nan Lu et al.

Figure 9. Vehicle makes right turn test. Figure 10. Vehicle makes left turn test.

to make a safety turn, it instantly lows the vehicle speed

to bring the vehicle safely back to its normal path and it
then accelerates again to make its turn. The fifth section
is the result of normal right turn, the vehicle decelerates
first before making turn and it accelerates again after
making 2/3 of the turn. In sixth section of Figure 10 it il-
lustrates the vehicle makes a dangerous left turn, the
speed is high when it makes left turn and it does not in-
tend to make any speed deceleration that it is easy to in-
cur accidence. In Figure 11 it is the situation when the
driver is awake but he makes an abrupt acceleration and
then makes an abrupt braking, his speed is accelerated to
certain level but he is afraid to make further acceleration
so that in it shows in the figure that the speed is keeping
at the same level for a while, it makes a little skid toward
the left when he abruptly applies the brake and all non-
fixed objects in the car are all getting loose and scattering
over everywhere.
4.3 Test Data and Data Analysis
From the data that we collected from the model of
vehicle moving behavior as described in the previous
section we conclude in Figure 12 the correlations be-
tween vehicle moving directions versus vehicle mov-
ing speeds. From this figure it can identify which is the
normal driving behavior that will not incur any dan-
gerous action and which is the driving behavior that
may easily generate accidence. It also displays their
distributions.
The test results as shown in Figures 13 and 14 are the
results possibly displayed in the users interface. As
shown in Figure 13, the behavior designated with green
line is classified as a normal condition while the danger-
ous behavior is displayed in red.
Figure 11. Vehicle abruptly accelerating/abruptly applies bra-
king test.
Each behavior is also given a name to signify its con-
dition. In actual implementation this system will be inte-
grated with the vehicles data bus and to provide the real
time data for the vehicles control system and have the
control system to make proper decision.
4.4 The Integration with Vehicles Data Bus
The acceleration impulse pre-warning system and
the driving behavior monitoring system as discussed in
the paper are the test systems designed only for experi-
mental purpose. However if they only possess these de-
scribed functions or they are separately implemented
they will not provide many practical usages. The most
important issue is that they need to be cooperated with
the vehicles data bus system so that it can convey the
measured data through the data bus and integrating to-
gether with the control system to accomplish the main
task of protecting the safeties of passengers lives and
properties.
 Intelligent Safety Warning and Alert System for Car Driving 403

Figure 12. Correlations between vehicle moving directions with vehicle moving speeds.

Figure 13. Model of abrupt acceleration.

5. Conclusion celerate the vehicle speed prior to the happening of acci-

If many instruments are equipped in a high speed
vehicle to provide more than necessary information to
the driver they will draw the driver to pay particular at-
tention to these instruments and furthermore he needs to
learn how to maneuver there instruments. Diversion, fa-
tigue are two fatal factors in the car accidents.
The main task of our study is trying through the study
of drivers driving behavior and in coordinating with the
information provided from the pre-warning system to de-
dent and if accident happens to reduce the damage to the
least level. In our study we also realize that in addition to
the items we have tested and verified they still have some
tasks we can study in the near future such as the task when
the car is running along a sloppy road how the driving be-
havior monitoring system can be incorporated with vehi-
cle speed data, which is transmitted and obtained through
the data bus system, to provide relevant data to the moni-
toring system to reflect a more accurate vehicles moving
condition and environment. To provide a more concise
404 Shih-Nan Lu et al.
Figure 14. Model of abnormal left turn.
format for the data transmitted in the data bus system to
save some system memory is also a challenge task.
References
[1] Ioannou, P., Xu, Z., Eckert, S. and Clemons, D. Sieja,
Intelligent Cruise Control: Theory and Experiment,
Decision and Control, 1993, Proceedings of the 32nd
IEEE Conference on, Vol. 2, pp. 1885-1890 (1993).
[2] Juanid, K. M., Wang, S., Usman, K. and Tao, W., In-
telligent Longitudinal Cruise Control by Quadratic
Minimization and Robust Synthesis, Vehicular Elec-
tronics and Safety, 2005, IEEE International Confer-
ence on, pp. 182-187 (2005).
[3] Mayr, R., Intelligent Cruise Control for Vehicles
Based on Feedback Linearization, American Control
Conference, Vol. 1, pp. 16-20 (1994).
[4] Sun, J. and Yang, Q., Modeling and Intelligent Con-
trol of Vehicle Active Suspension System, Robotics,
Automation and Mechatronics, 2008, IEEE Confer-
ence on, pp. 239-242 (2008).
[5] Mayr, R. and Bauer, Q., Safety Issues in Intelligent
Cruise Control, Intelligent Transportation Systems,
1999, Proceedings, 1999 IEEE/IEEJ/JSAI Interna-
tional Conference on, pp. 970-975 (1999).
[6] Hahn, S., Automation of Driving Functions-Future
Development, Benefits and Pitfalls, Intelligent Vehi-
cles Symposium, 1996, Proceedings of the 1996 IEEE,
pp. 309-312 (1999).
[7] Kim, M. H., Lee, Y. T. and Son, J., Age-Related
Physical and Emotional Characteristics to Safety
Warning Sounds: Design Guidelines for Intelligent
Vehicles, IEEE Transactions on Systems, Man, and
Cybernetics, Part C: Applications and Reviews, Vol.
99, pp. 1-7 (2010).
[8] Ruan, J., Yang, F., Song, R. and Li, Y., Study on
ADRC-Based Intelligent Vehicle Lateral Locomotion
Control, Intelligent Control and Automation, 2008.
WCICA 2008. 7th World Congress on, pp. 2619-2624
(2008).
[9] Thomas Kibler, Stefan Poferl, Gotthard Bck, Hans-
Peter Huber and Eberhard Zeeb, Optical Data Buses
for Automotive Applications, Journal of Lightwave
Technology, Vol. 22 (2004).
[10] MOST MOST Cooperation Official, 2008 MOST
Cooperation, http://www.mostcooperation.com.
[11] Automotive Buses, Automotive Buses and Vehicle
Buses, http://www.interfacebus.com/Design_Connec-
tor_Automotive.html.
[12] D. John Oliver, Implementing the J1850 Protocol,
Intel Corporation.
[13] CAN in Automation (CiA), Controller Area Network
(CAN) - Protocol, http://www.can-cia.org/can/pro-
tocol/.
[14] BOSCH, CAN Homepage, http://www.can.bosch.
com.
Manuscript Received: Oct. 10, 2008
Accepted: May 25, 2010